J-20 5th Gen Fighter Thread VIII

[latenlazy]

Yeah and i asked when the surface is curved am i supposed to break it into a box with mm width ? and is there any academic source ?

Yes? That is literally what I said *at the very start* of this exchange...

If you ever write a thesis, you cannot answer with "it's basic science in school".

I have never seen a science paper that has needed to source the claim that "we cannot measure accurately this thing we're interested in because our instruments don't have enough sensitivity", and I have read a whole lot of science papers (where this kind of commentary is frequently brought up).

 

[Stealthflanker]

Yes? That is literally what I said *at the very start* of this exchange...

And i am asking for a source too, you dont provide. While other researchers in RCS doesnt seem to do what you said. e.g Sukharevsky.

I have never seen a science paper that has needed to source the claim that "we cannot measure accurately this thing we're interested because our instruments don't have enough sensitivity", and I have read a whole lot of science papers (where this kind of commentary is frequently brought up).

Well i am asking for a paper that mentions the kind of poly dimensions. Like say how many polycounts. or how big they should be. you dont provide that, why ?

 

[latenlazy]

And i am asking for a source too, you dont provide. While other researchers in RCS doesnt seem to do what you said. e.g Sukharevsky.




Well i am asking for a paper that mentions the kind of poly dimensions. Like say how many polycounts. or how big they should be. you dont provide that, why ?

Please, Log in or Register to view URLs content!

If you can understand this paper then you would realize why the point I’m making is so elementary it’s nonsensical to ask for sources.

Please, Log in or Register to view URLs content!

Here’s another paper on RCS modeling for complex objects. If you understand what they’re doing with their object shape modeling to get more accurate RF interactions in their model you would understand why rough polynomial approximations of an object (which is what your models have depended on) is a problematic and inaccurate approach.

 

[Stealthflanker]

Here’s another paper on RCS modeling for complex objects. If you understand what they’re doing with their object shape modeling to get more accurate RF interactions in their model you would understand why rough polynomial approximations of an object (which is what your models have depended on) is a problematic and inaccurate approach.

Well i understand. They are using NURBS which basically STEP file. and they said it's alternative not substitute. Nor it said requires smaller surface or "resolution"

Your second paper doesn't even talk about modeling resolution at all nor even in the microwave range it talks about compact RCS range using extremely short wavelength for measurement. Are you really reading them ?
----------

Well TBH This is the kind of answer i wish you could provide... This is from POFACET's manual.





Look at the Rule of Thumb section.
-Object must be large which my model is.
-Facet's edge which basically the facet's dimension.. it has to be bigger than the wavelength.. not smaller apparently.

But then ANSYS is SBR+ solver while POFACETS is PO. Will the rule of thumb works ?

That's how a clear and concise explanation is.

Also i apologize for being rude and i actually forget to thank you for your care.

 

[latenlazy]

Well i understand. They are using NURBS which basically STEP file. and they said it's alternative not substitute. Nor it said requires smaller surface or "resolution"

A better alternative is a substitute. NURBS is an attempt to deal with curvature issues *because* model resolution is a problem.

Your second paper doesn't even talk about modeling resolution at all nor even in the microwave range it talks about compact RCS range using extremely short wavelength for measurement. Are you really reading them ?
----------

...Resolution=feature scaling relative to factors.

Well TBH This is the kind of answer i wish you could provide... This is from POFACET's manual.


View attachment 116304


Look at the Rule of Thumb section.
-Object must be large which my model is.
-Facet's edge which basically the facet's dimension.. it has to be bigger than the wavelength.. not smaller apparently.

But then ANSYS is SBR+ solver while POFACETS is PO. Will the rule of thumb works ?

That's how a clear and concise explanation is.

1) That's for double *curved* surfaces. 2) Your model has *0* curvature.

 

[Stealthflanker]

A better alternative is a substitute. NURBS is an attempt to deal with curvature issues *because* model resolution is a problem.

Well facets are still used tho. and not resolution. They are talking about reduction of computation requirement and time. I mean dont you see the table at the conclusion page ? They list time to simulate.

...Resolution=feature scaling relative to factors.

No. It talks about Scale model RCS testing using laser.

You dont have to test a full size model everytime in real range. But you can test a scaled down/smaller model. But there is rule of thumb for this which you also need to increase the frequency of your measurement. That's what the paper talk about. They measure RCS with laser on a small scale model.

1) That's for double *curved* surfaces. 2) Your model has *0* curvature.

1. Aircraft will have plenty of those.
2. It has except you complained that i dont make the facets small enough and when i ask for source on how small they should be, you make this long exchange.

This is from Sukharevsky's book btw



Does this model looks like anything you imagine ? And second one.



The book doesnt list however how small the facets it is.

 

[latenlazy]

Well facets are still used tho. and not resolution. They are talking about reduction of computation requirement and time. I mean dont you see the table at the conclusion page ? They list time to simulate.

NURBS simulates curvature from the faceted wireframe.

No. It talks about Scale model RCS testing using laser.

No one is minding appropriate model scale relative to wavelength scale if resolution isn't essential to getting accurate modeling.

You dont have to test a full size model everytime in real range. But you can test a scaled down/smaller model. But there is rule of thumb for this which you also need to increase the frequency of your measurement. That's what the paper talk about. They measure RCS with laser on a small scale model.

Yes *because resolution scaled appropriately to wavelength sizes is essential to getting accurate modeling*.

1. Aircraft will have plenty of those.
2. It has except you complained that i dont make the facets small enough and when i ask for source on how small they should be, you make this long exchange.

This is from Sukharevsky's book btw

View attachment 116305

Does this model looks like anything you imagine ? And second one.

This model has curvature. Yours does not. Again, *the issue here is that you are using a very coarse polygonal model*.

EDIT: Anyways, last I am going to say about this because I think this exchange is starting to derail the thread. We can let everyone else judge whether your modeling exercises meets sufficient rigor to generate meaningful data.

 

[Stealthflanker]

NURBS simulates curvature from the faceted wireframe.

So you agree that i have to use STEP then ? But Step files doesnt have the kind of "resolution" you imagine.

No one is minding appropriate model scale relative to wavelength scale if resolution isn't essential to getting accurate modeling.

Are you really reading the paper you provided yourself ? and are you kidding ? There is a whole chapter of scaling laws in E.Knott's book "Radar Cross Section Measurements"

This model has curvature. Yours does not. Again, *the issue here is that you are using a very coarse polygonal model*.

Well it has, and your problem is wanting to make it smaller yet unable to provide backing source on why. Maybe we can agree to disagree here finally.

 

[latenlazy]

So you agree that i have to use STEP then ? But Step files doesnt have the kind of "resolution" you imagine.

*facepalm* You need higher resolution if you are going to use a model that doesn't factor in curvature. If your model approximates curvature then the problems aren't as serious (but they don't go away). I would caution though that even if you use STEP there's no guarantee that the results will be representatively accurate *for comparison between two different designs*.

Are you really reading the paper you provided yourself ? and are you kidding ? There is a whole chapter of scaling laws in E.Knott's book "Radar Cross Section Measurements"

There wouldn't be scaling laws if feature size relative to beam wavelength wasn't essential to RCS.

Well it has, and your problem is wanting to make it smaller yet unable to provide backing source on why. Maybe we can agree to disagree here finally.

Do you think no one can see the RCS output from your own models...your model does not have curvature.

Honestly, if model resolution weren't a problem you wouldn't have mentioned needing to redo the F-35 model with higher polynomial counts. *You yourself are implicitly admitting that your models are too coarse and that model resolution matters when you acknowledge your F-35 model needs greater polycounts*.

 

[latenlazy]

Forgot to add this in my earlier reply but just for general public education and clarification of the contentions, for those wondering why adding curvature approximation to a physical model reduces problems from resolution limitations, curvature essentially represents infinite polygonal segmentation of a shape. When you are increasing polygon counts and thus reducing polygon features sizes *what you are doing is trying to do closer approximations of curvature*. In other words curvature approximates infinite resolution, which is why if he adopted a curvature approximation method into his 3D model then polygon count wouldn’t be as much of an issue (but given curvature approximation is still an approximation residual errors can still potentially overwhelm significance). Anyone who understands the basics of geometric modeling would know this. If stealthflanker’s 3D models had curvature approximations in them he wouldn’t be worried about whether his F-35 model had sufficient polygon counts.